Projet Kit Data Science

EL HALABI Jad Rayan - MS Big Data

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.patches import Patch
import seaborn as sns
import requests
import os
import re
from bs4 import BeautifulSoup
import datetime
import xlwings
import locale
import random
import glob
import plotly.graph_objects as go
import plotly.express as px
from plotly.offline import init_notebook_mode, iplot

from sklearn.model_selection import train_test_split
from sklearn import linear_model, preprocessing, cluster
from sklearn.metrics import r2_score, mean_squared_error
from statsmodels.tsa.holtwinters import SimpleExpSmoothing
from statsmodels.tsa.seasonal import seasonal_decompose
from statsmodels.tsa.statespace.sarimax import SARIMAX

Récupération des données

Attention : le code ci-dessous permet de récupérer les données mais cette étape prend du temps. Le dataframe a donc été sauvegardé en fichier pickle

# we request the url
req = requests.get('https://www.vendeeglobe.org/fr/classement')

# we get our soup from the downloaded content
soup = BeautifulSoup(req.content)

dates = []
# we extract the arrival dates from the website
for td in soup.find('table', attrs={'class': 'ranking-report'}).findAll('td', attrs={'class': 'row-layout', 'colspan': '2'}):
    dates.append(datetime.datetime.strptime(td.text, 'Date arrivée : %d/%m/%Y %H:%M:%S'))

# we extract the first arrival
first = min(dates)
# we extract the available values from the website's list
for option in soup.find('select', attrs={'class': 'form__input m--select onsubmit-rank'}).findAll('option'):
    date = option.attrs['value']
    # we skip if the date is empty or after the first arrival
    if date != '' and datetime.datetime.strptime(date, '%Y%m%d_%H%M%S') < first:
        file = requests.get('https://www.vendeeglobe.org/download-race-data/vendeeglobe_'+date+'.xlsx')
        open('data/'+date+'.xlsx', 'wb').write(file.content)

def xlsx_broken(file):
    excel_app = xlwings.App(visible=False)
    excel_book = excel_app.books.open(file)
    excel_book.save()
    excel_book.close()
    excel_app.quit()
    return True

for f in glob.glob("data/*"):
    xlsx_broken(f)

files = [f for f in os.listdir('data/') if not f.startswith('.')][1:-1]

# we load a file
def load_data(filename):
    df = pd.read_excel('data/'+filename, usecols='B:U', skiprows=[0,1,2,3],
                    names=['Rank', 'Nat. Sail', 'Skipper/crew', 'Hour FR', 'Latitude', 'Longitude',
                    'Heading ° 30min', 'Speed kts 30min', 'VMG kts 30min', 'Distance nm 30min',
                    'Heading ° last report', 'Speed kts last report', 'VMG kts last report','Distance nm last report',
                    'Heading ° 24h', 'Speed kts 24h', 'VMG kts 24h', 'Distance nm 24h', 'DTF', 'DTL'])

    # we drop the na values
    df = df.dropna()

    # we split the string values into separate columns
    df['Nationality'] = df['Nat. Sail'].str.split(' ').str[0].str.replace('\n', '')
    df['Sail'] = df['Nat. Sail'].str.split(' ').str[1].apply(int)
    df = df.drop('Nat. Sail', axis=1)

    df['Skipper'] = df['Skipper/crew'].str.split('\n').str[0]
    df['Crew'] = df['Skipper/crew'].str.split('\n').str[1].str.title()
    df = df.drop('Skipper/crew', axis=1)
    
    # we convert the latitude and longitude into decimal
    # Decimal degrees = Degrees + (Minutes/60) + (Seconds/3600), N & E = + /  S & W = - 
    df['Latitude'] = df['Latitude'].apply(lambda s: (1 if s.split('°')[1].split('.')[1][-1] == 'N' else -1)
                                                    * (int(s.split('°')[0])
                                                    + int(s.split('°')[1].split('.')[0])/60
                                                    + int(re.sub('[^0-9]', '', s.split('°')[1].split('.')[1]))/3600))
    df['Longitude'] = df['Longitude'].apply(lambda s: (1 if s.split('°')[1].split('.')[1][-1] == 'E' else -1)
                                                    * (int(s.split('°')[0])
                                                    + int(s.split('°')[1].split('.')[0])/60
                                                    + int(re.sub('[^0-9]', '', s.split('°')[1].split('.')[1]))/3600))

    # we convert the time into a proper datetime value
    date = re.search('([0-9]{8})', filename).group(0)
    df['Hour FR'] = df['Hour FR'].apply(lambda x: datetime.datetime.strptime(date+'_'+x[:5], '%Y%m%d_%H:%M'))

    # we set it as the index
    df = df.set_index('Hour FR')

    # we convert all the numerical values into floats
    df['Rank'] = df['Rank'].apply(int)
    numerical_cols = ['Heading ° 30min', 'Speed kts 30min', 'VMG kts 30min', 'Distance nm 30min',
                    'Heading ° last report', 'Speed kts last report', 'VMG kts last report','Distance nm last report',
                    'Heading ° 24h', 'Speed kts 24h', 'VMG kts 24h', 'Distance nm 24h', 'DTF', 'DTL']
    df[numerical_cols] = df[numerical_cols].apply(lambda x: x.str.replace(re.compile('[a-z]+|°'), '')).apply(pd.to_numeric)

    # we reorder the columns
    df = df[['Rank', 'Nationality', 'Sail', 'Skipper', 'Crew', 'Latitude', 'Longitude', *numerical_cols]]

    return df

# For example let's read the first file
df = load_data(files[1])
df.head()

	Rank	Nationality	Sail	Skipper	Crew	Latitude	Longitude	Heading ° 30min	Speed kts 30min	VMG kts 30min	...	Heading ° last report	Speed kts last report	VMG kts last report	Distance nm last report	Heading ° 24h	Speed kts 24h	VMG kts 24h	Distance nm 24h	DTF	DTL
Hour FR
2020-11-08 17:00:00	1	FRA	8	Jérémie Beyou	Charal	46.353333	-2.597222	268	26.0	22.6	...	262	21.1	19.7	31.7	254	1.4	1.4	34.5	24265.1	0.0
2020-11-08 17:00:00	2	FRA	79	Charlie Dalin	Apivia	46.296111	-2.529167	262	24.8	22.8	...	357	0.0	0.0	2782.5	247	1.4	1.4	32.9	24265.8	0.7
2020-11-08 17:00:00	3	GBR	99	Alex Thomson	Hugo Boss	46.318889	-2.535000	263	22.8	20.8	...	257	19.9	19.1	29.9	249	1.4	1.3	32.7	24266.3	1.2
2020-11-08 17:00:00	4	FRA	59	Thomas Ruyant	Linkedout	46.329167	-2.541389	266	24.3	21.6	...	260	19.9	18.9	29.9	251	1.4	1.3	32.7	24266.3	1.2
2020-11-08 17:00:00	5	MON	10	Boris Herrmann	Seaexplorer - Yacht Club De Monaco	46.306111	-2.500833	263	23.9	21.9	...	257	19.0	18.2	28.2	248	1.3	1.3	31.5	24267.3	2.2

5 rows × 21 columns

df_list = []
for f in files:
    df_list.append(load_data(f))
df_all_rankings = pd.concat(df_list)
df_all_rankings.head()

	Rank	Nationality	Sail	Skipper	Crew	Latitude	Longitude	Heading ° 30min	Speed kts 30min	VMG kts 30min	...	Heading ° last report	Speed kts last report	VMG kts last report	Distance nm last report	Heading ° 24h	Speed kts 24h	VMG kts 24h	Distance nm 24h	DTF	DTL
Hour FR
2020-11-08 15:30:00	1	FRA	18	Louis Burton	Bureau Vallée 2	46.412778	-1.846667	241	17.7	17.5	...	357	0.0	0.0	2788.0	201	0.3	0.3	6.1	24293.9	0.0
2020-11-08 15:31:00	2	MON	10	Boris Herrmann	Seaexplorer - Yacht Club De Monaco	46.409444	-1.839444	241	11.1	10.9	...	357	0.0	0.0	2787.9	196	0.3	0.2	6.0	24294.2	0.4
2020-11-08 15:30:00	3	FRA	8	Jérémie Beyou	Charal	46.425278	-1.844167	244	15.5	15.5	...	357	0.0	0.0	2788.5	199	0.2	0.2	5.5	24294.3	0.5
2020-11-08 15:30:00	4	FRA	59	Thomas Ruyant	Linkedout	46.419722	-1.835556	244	13.2	13.1	...	357	0.0	0.0	2788.3	196	0.2	0.2	5.6	24294.5	0.6
2020-11-08 15:30:00	5	FRA	53	Maxime Sorel	V And B Mayenne	46.416389	-1.832222	246	10.9	10.9	...	357	0.0	0.0	2788.1	195	0.8	0.7	5.8	24294.5	0.6

5 rows × 21 columns

# we request the url
headers = {'User-Agent' : 'Mozilla/5.0'}
req_skip = requests.get('https://www.vendeeglobe.org/fr/glossaire', headers = headers)

# we raise an Exception if the status code is not 200
if req_skip.status_code != 200:
    raise Exception('Couldn\'t get open URL')

# we get our soup from the downloaded content
soup_skip = BeautifulSoup(req_skip.content)

# we extract the arrival dates from the website
df_list = []
for div in soup_skip.findAll('div', attrs={'class': 'boats-list__popup-infos'}):
    # we get the skipper name
    skip_name = div.find('h3', attrs={'class': 'boats-list__popup-title'}).text.title()
    skippers_info = {'Equipage':skip_name}
    # we get the rest of the data
    for li in div.find('ul', attrs={'class': 'boats-list__popup-specs-list'}).findAll('li'):
        info = li.text.split(' : ')
        skippers_info[info[0].strip()] = [info[1].strip()]

    # we append all data as dataframes
    df_list.append(pd.DataFrame(skippers_info))

# we generate a dataframe containing all of the extracted data
df_skippers = pd.concat(df_list, ignore_index=True)

df_skippers.head()

# we replace the former names of the boat by a propoer list of strings
df_skippers['Anciens noms du bateau'] = df_skippers['Anciens noms du bateau'].fillna('').str.strip().str.split(',')

# we convert all numerical value into floats or int
numerical_cols = ['Longueur', 'Largeur', 'Tirant d\'eau', 'Déplacement (poids)','Hauteur mât','Surface de voiles au près','Surface de voiles au portant']
df_skippers[numerical_cols] = df_skippers[numerical_cols].apply(lambda x: x.str.replace(re.compile('m2|m²|m|t[a-z]*|nc|NC'), '').str.replace(',', '.')).apply(pd.to_numeric)

# we rename the columns to include the units
df_skippers = df_skippers.rename(columns={'Longueur': 'Longueur (m)', 'Largeur': 'Largeur (m)', 'Tirant d\'eau': 'Tirant d\'eau (m)', 'Déplacement (poids)': 'Déplacement (poids) (t)','Hauteur mât': 'Hauteur mât (m)','Surface de voiles au près': 'Surface de voiles au près (m2)', 'Surface de voiles au portant': 'Surface de voiles au portant (m2)'})

# we correct the 'foiler' vs 'foils' categories
df_skippers['Nombre de dérives'] = df_skippers['Nombre de dérives'].str.replace('foiler', 'foils')

# we convert all dates into datetimes
locale.setlocale(locale.LC_ALL, 'fr')
df_skippers['Date de lancement'] = df_skippers['Date de lancement'].apply(lambda x: datetime.datetime.strptime(x, '%d %B %Y'))

df_skippers.head()

Sauvegarde du dataframe en pickle

df_all_rankings.to_pickle('df_all_rankings.pkl')
df_skippers.to_pickle('df_skippers.pkl')

Chargement du pickle, les cellules suivantes peuvent être executées

df_all_rankings = pd.read_pickle('df_all_rankings.pkl')
df_skippers = pd.read_pickle('df_skippers.pkl')

# we merge the two dataframes
df_ranking_skippers = pd.merge(left=df_all_rankings.reset_index(), right=df_skippers, left_on='Crew', right_on='Equipage', how='inner').set_index('Hour FR')
df_ranking_skippers.sample(5)

	Rank	Nationality	Sail	Skipper	Crew	Latitude	Longitude	Heading ° 30min	Speed kts 30min	VMG kts 30min	...	Date de lancement	Longueur (m)	Largeur (m)	Tirant d'eau (m)	Déplacement (poids) (t)	Nombre de dérives	Hauteur mât (m)	Voile quille	Surface de voiles au près (m2)	Surface de voiles au portant (m2)
Hour FR
2021-01-23 11:30:00	7	ITA	34	Giancarlo Pedote	Prysmian Group	31.225278	-32.255278	44	8.3	8.3	...	2015-04-02	18.28	5.80	4.5	8.0	foils	29.0	acier forgé	300	600
2020-12-09 14:30:00	2	FRA	59	Thomas Ruyant	Linkedout	-38.726667	81.827500	132	15.7	15.2	...	2019-09-03	18.28	5.85	4.5	8.0	foils	29.0	acier forgé	350	560
2021-01-27 14:30:00	21	FRA	71	Manuel Cousin	Groupe Sétin	-24.011667	-26.806389	347	8.9	7.9	...	2007-02-02	18.28	5.80	4.5	9.0	2 asymétriques	28.5	basculante sur vérin hydraulique	270	560
2021-01-22 04:30:00	17	GBR	777	Pip Hare	Medallia	-26.868611	-36.178333	352	4.2	4.0	...	1999-07-03	18.28	5.70	4.5	9.0	2	29.0	carbone	300	580
2021-02-25 11:30:00	25	FIN	222	Ari Huusela	Stark	34.818056	-33.468333	40	12.1	11.7	...	2007-08-06	18.28	5.80	4.5	8.5	2	28.0	acier	270	580

5 rows × 36 columns

Exploration des données

df_all_rankings.shape

(15272, 21)

print("les équipages, le nom et numéro de leur bateau, leur nationalité ainsi que la distance moyenne parcourue par jour")
df_all_rankings.groupby(['Sail', 'Nationality', 'Skipper', 'Crew'])['Distance nm 24h'].mean().sort_values(ascending=True)

les équipages, le nom et numéro de leur bateau, leur nationalité ainsi que la distance moyenne parcourue par jour

Sail  Nationality  Skipper               Crew                                  
69    FRFRA        Sébastien Destremau   Merci                                     212.909353
222   FIN          Ari Huusela           Stark                                     234.952211
56    FRA          Fabrice Amedeo        Newrest - Art Et Fenetres                 236.703500
72    FRA          Alexia Barrier        Tse - 4Myplanet                           238.259403
50    FRA          Miranda Merron        Campagne De France                        261.288670
71    FRA          Manuel Cousin         Groupe Sétin                              262.840675
83    FRA          Clément Giraud        Compagnie Du Lit - Jiliti                 264.040067
6     FRA          Nicolas Troussel      Corum L'Épargne                           269.396000
33    ESP          Didac Costa           One Planet One Ocean                      277.675171
777   GBR          Pip Hare              Medallia                                  279.559582
7     SUI          Alan Roura            La Fabrique                               284.864510
92    FRA          Stéphane Le Diraison  Time For Oceans                           286.359860
14    FRA          Arnaud Boissieres     La Mie Câline - Artisans Artipôle         287.263604
11    JPN          Kojiro Shiraishi      Dmg Mori Global One                       288.328196
49    FRA          Romain Attanasio      Pure - Best Western Hotels And Resorts    289.994085
30    FRA          Clarisse Cremer       Banque Populaire X                        302.536711
99    GBR          Alex Thomson          Hugo Boss                                 304.442949
109   FRA          Samantha Davies       Initiatives - Coeur                       306.887117
8     FRA          Jérémie Beyou         Charal                                    309.184045
2     FRFRA        Armel Tripon          L'Occitane En Provence                    312.056496
53    FRA          Maxime Sorel          V And B Mayenne                           313.839919
4     FRA          Sébastien Simon       Arkea Paprec                              314.924684
27    FRA          Isabelle Joschke      Macsf                                     318.225266
1     FRA          Jean Le Cam           Yes We Cam !                              320.729303
1000  FRA          Damien Seguin         Groupe Apicil                             320.936420
9     FRA          Benjamin Dutreux      Omia - Water Family                       322.044196
34    ITA          Giancarlo Pedote      Prysmian Group                            325.390329
85    FRA          Kevin Escoffier       Prb                                       327.411679
17    FRA          Yannick Bestaven      Maître Coq Iv                             327.464803
10    MON          Boris Herrmann        Seaexplorer - Yacht Club De Monaco        327.923093
18    FRA          Louis Burton          Bureau Vallée 2                           330.648447
59    FRA          Thomas Ruyant         Linkedout                                 330.988017
79    FRA          Charlie Dalin         Apivia                                    331.893971
Name: Distance nm 24h, dtype: float64

print("classement moyen de chaque équipage")
df_all_rankings.groupby(['Nationality', 'Skipper', 'Crew'])['Rank'].mean().sort_values()

classement moyen de chaque équipage

Nationality  Skipper               Crew                                  
FRA          Charlie Dalin         Apivia                                     2.286902
             Thomas Ruyant         Linkedout                                  3.524793
             Yannick Bestaven      Maître Coq Iv                              4.374741
             Jean Le Cam           Yes We Cam !                               5.629098
             Kevin Escoffier       Prb                                        5.846715
             Damien Seguin         Groupe Apicil                              6.382716
             Louis Burton          Bureau Vallée 2                            6.809524
MON          Boris Herrmann        Seaexplorer - Yacht Club De Monaco         6.921649
FRA          Nicolas Troussel      Corum L'Épargne                            7.260000
GBR          Alex Thomson          Hugo Boss                                  7.461538
FRA          Benjamin Dutreux      Omia - Water Family                        7.572301
             Samantha Davies       Initiatives - Coeur                       10.337423
             Sébastien Simon       Arkea Paprec                              10.398734
ITA          Giancarlo Pedote      Prysmian Group                            10.423868
FRA          Maxime Sorel          V And B Mayenne                           10.659274
             Isabelle Joschke      Macsf                                     11.510638
             Romain Attanasio      Pure - Best Western Hotels And Resorts    13.415896
             Clarisse Cremer       Banque Populaire X                        13.508604
FRFRA        Armel Tripon          L'Occitane En Provence                    16.076772
SUI          Alan Roura            La Fabrique                               16.466783
FRA          Arnaud Boissieres     La Mie Câline - Artisans Artipôle         17.452297
             Stéphane Le Diraison  Time For Oceans                           17.818499
GBR          Pip Hare              Medallia                                  19.221254
ESP          Didac Costa           One Planet One Ocean                      19.964041
FRA          Manuel Cousin         Groupe Sétin                              21.302251
             Jérémie Beyou         Charal                                    21.430427
JPN          Kojiro Shiraishi      Dmg Mori Global One                       22.084063
FRA          Miranda Merron        Campagne De France                        23.522167
             Clément Giraud        Compagnie Du Lit - Jiliti                 24.295492
             Alexia Barrier        Tse - 4Myplanet                           24.819403
FIN          Ari Huusela           Stark                                     25.955777
FRA          Fabrice Amedeo        Newrest - Art Et Fenetres                 26.965000
FRFRA        Sébastien Destremau   Merci                                     27.872902
Name: Rank, dtype: float64

Distances parcourues par 5 skippers aléatoires

# cumulative distance for 5 random skippers
Skippers = np.random.choice(df_all_rankings['Skipper'].unique(), 5)

for skipper in Skippers[:5]:
    skip = df_all_rankings.loc[(df_all_rankings['Skipper'] == skipper),"Distance nm last report"]
    skip.cumsum().plot(label = skipper)
    
plt.ylabel('Distance totale parcourue (km)')
plt.legend();

fig = go.Figure(go.Scattermapbox())
i=0
for skipper in Skippers: 
    sk = df_all_rankings.loc[df_all_rankings['Skipper'] == skipper]
    stamp = sk.index

    fig.add_trace(go.Scattermapbox(
        
        mode = "markers+lines",
        lon = sk["Longitude"],
        lat = sk["Latitude"],
        name = skipper,
        marker = {'size': 3},
                text= stamp ))
    i+=1
     

fig.update_layout(
    margin ={'l':0,'t':0,'b':0,'r':0},
    mapbox = {
        'center': {'lon': 10, 'lat': 10},
        'style': "stamen-terrain",
        'center': {'lon': -20, 'lat': -20},
        'zoom': 1})

fig.show()

Relation VMG et Classement

# we prepare our plot
fig = plt.figure(figsize=(20, 8));
k = 1

# we pick 4 rndom skippers
skippers = np.random.choice(df_all_rankings['Skipper'].unique(), 4)
columns_graph = ['Speed kts 24h', 'VMG kts 24h', 'Distance nm 24h']

# for each column
for col in columns_graph:
    # we add a subplot
    ax = fig.add_subplot(1, 3, k)
    # we plot the categories in each subplot
    for sk in skippers:
        ax.plot(df_all_rankings[df_all_rankings['Skipper'] == sk].index,
                df_all_rankings.loc[df_all_rankings['Skipper'] == sk, col],
                label = sk, alpha=0.7);
        ax.set_title(col);
    ax.legend();
    k+=1
fig.tight_layout()

# we get the average 24h VMG and Distance for the crew we just isolated
VMG_24_APIVIA = df_all_rankings.loc[df_all_rankings['Crew'] == crew_first_last, ('VMG kts 24h', 'Rank')]

# we graph its regression using Seaborn
sns.set(rc={'figure.figsize':(11.7,8.27)})
sns.regplot(x='VMG kts 24h', y='Rank', data=VMG_24_APIVIA,  marker="+");

Exploitation des données sur les voiliés

# we agregate some numerical values by Type of foil
df_ranking_skippers.groupby(['Nombre de dérives']).agg({'Crew': lambda x: x.nunique(),
                                                        'Speed kts 24h': np.mean,
                                                        'VMG kts 24h': np.mean,
                                                        'Distance nm 24h': np.mean})

	Crew	Speed kts 24h	VMG kts 24h	Distance nm 24h
Nombre de dérives
2	7	11.234977	10.182537	269.502003
2 asymétriques	1	10.951286	9.707395	262.840675
foils	15	12.906236	11.636471	309.714768

sns.barplot(data=df_ranking_skippers, y='Crew', x='Distance nm 24h', hue='Nombre de dérives', orient='h');

impact du foil

pl = df_ranking_skippers.groupby('Skipper')[['VMG kts 24h', 'Speed kts 24h']].mean()
pl['Foils'] = df_ranking_skippers.groupby('Skipper')['Nombre de dérives'].unique().apply(pd.Series)

pl.head()

	VMG kts 24h	Speed kts 24h	Foils
Skipper
Alan Roura	10.641608	11.870455	foils
Alex Thomson	10.966667	12.698718	foils
Ari Huusela	8.673752	9.788873	2
Armel Tripon	11.930118	13.003150	foils
Arnaud Boissieres	10.724558	11.969788	foils

px.pie(pl, values='VMG kts 24h', names='Foils', title='VMG split by foils')

# we prepare our plot
fig = plt.figure(figsize=(20, 8));
k = 1

# we extract the 4 different types of foils
types = df_ranking_skippers['Nombre de dérives'].unique()
columns_graph = ['Speed kts 24h', 'VMG kts 24h', 'Distance nm 24h']

# for each column
for col in columns_graph:
    # we add a subplot
    ax = fig.add_subplot(1, 3, k)
    # we plot the categories in each subplot
    for t in types:
        ax.plot(df_ranking_skippers[df_ranking_skippers['Nombre de dérives'] == t].resample('D').mean().index,
                df_ranking_skippers.loc[df_ranking_skippers['Nombre de dérives'] == t, col].resample('D').mean(),
                label = t, alpha=0.7);
        ax.set_title(col);
    ax.legend();
    k+=1
fig.tight_layout()

print("les voiliers équipés de foils ont des meilleurs résultats (en terme de vitesse, VMG et distance moyennes sur 24h) que ceux équipés de 2 dérives")

les voiliers équipés de foils ont des meilleurs résultats (en terme de vitesse, VMG et distance moyennes sur 24h) que ceux équipés de 2 dérives

distance parcourue par voilier, avec et sans foil

# we prepare our plot
fig = plt.figure(figsize=(20, 8));
k = 1

# we pick 4 random crews
crew = df_all_rankings['Crew'].unique()
columns_graph = ['Distance nm 30min', 'Distance nm 24h', 'Distance nm last report']

# we create an aggregate, that will be useful to generate a colors map
agg_derive = df_ranking_skippers.pivot_table(values='Nombre de dérives', index='Crew', aggfunc='first')
# we generate the colors map
colors = ['r' if nd=='foils' else 'g' for nd in agg_derive.values]

# we create an aggregate for the graph
agg_crew = df_all_rankings.pivot_table(values=columns_graph, index='Crew', aggfunc='mean')

# we generate custom legends
custom_legend = [Patch(color='r', label='foils'),
                Patch(color='g', label='others/unknown')]

# for each column
for col in columns_graph:
    # we add a subplot
    ax = fig.add_subplot(1, 3, k)
    # we plot the columns values in each subplot
    ax.barh(y=agg_crew.index, width=agg_crew[col], color=colors);
    ax.set_title(col);

    # we hide the labels for the other 2 plots
    if k > 1:
        ax.get_yaxis().set_visible(False)
    k+=1
    # we generate and attach the custom legend
    ax.legend(handles=custom_legend)
fig.tight_layout()

print("A nouveau, on voit que les voiliers avec foil sont plus performants")

A nouveau, on voit que les voiliers avec foil sont plus performants

! jupyter nbconvert --to html "EL HALABI_KitDataScience_Projet Final.ipynb"

This application is used to convert notebook files (*.ipynb) to various other
formats.

WARNING: THE COMMANDLINE INTERFACE MAY CHANGE IN FUTURE RELEASES.

Options
=======
The options below are convenience aliases to configurable class-options,
as listed in the "Equivalent to" description-line of the aliases.
To see all configurable class-options for some <cmd>, use:
    <cmd> --help-all

--debug
    set log level to logging.DEBUG (maximize logging output)
    Equivalent to: [--Application.log_level=10]
--generate-config
    generate default config file
    Equivalent to: [--JupyterApp.generate_config=True]
-y
    Answer yes to any questions instead of prompting.
    Equivalent to: [--JupyterApp.answer_yes=True]
--execute
    Execute the notebook prior to export.
    Equivalent to: [--ExecutePreprocessor.enabled=True]
--allow-errors
    Continue notebook execution even if one of the cells throws an error and include the error message in the cell output (the default behaviour is to abort conversion). This flag is only relevant if '--execute' was specified, too.
    Equivalent to: [--ExecutePreprocessor.allow_errors=True]
--stdin
    read a single notebook file from stdin. Write the resulting notebook with default basename 'notebook.*'
    Equivalent to: [--NbConvertApp.from_stdin=True]
--stdout
    Write notebook output to stdout instead of files.
    Equivalent to: [--NbConvertApp.writer_class=StdoutWriter]
--inplace
    Run nbconvert in place, overwriting the existing notebook (only 
    relevant when converting to notebook format)
    Equivalent to: [--NbConvertApp.use_output_suffix=False --NbConvertApp.export_format=notebook --FilesWriter.build_directory=]
--clear-output
    Clear output of current file and save in place, 
    overwriting the existing notebook.
    Equivalent to: [--NbConvertApp.use_output_suffix=False --NbConvertApp.export_format=notebook --FilesWriter.build_directory= --ClearOutputPreprocessor.enabled=True]
--no-prompt
    Exclude input and output prompts from converted document.
    Equivalent to: [--TemplateExporter.exclude_input_prompt=True --TemplateExporter.exclude_output_prompt=True]
--no-input
    Exclude input cells and output prompts from converted document. 
    This mode is ideal for generating code-free reports.
    Equivalent to: [--TemplateExporter.exclude_output_prompt=True --TemplateExporter.exclude_input=True]
--allow-chromium-download
    Whether to allow downloading chromium if no suitable version is found on the system.
    Equivalent to: [--WebPDFExporter.allow_chromium_download=True]
--log-level=<Enum>
    Set the log level by value or name.
    Choices: any of [0, 10, 20, 30, 40, 50, 'DEBUG', 'INFO', 'WARN', 'ERROR', 'CRITICAL']
    Default: 30
    Equivalent to: [--Application.log_level]
--config=<Unicode>
    Full path of a config file.
    Default: ''
    Equivalent to: [--JupyterApp.config_file]
--to=<Unicode>
    The export format to be used, either one of the built-in formats
    ['asciidoc', 'custom', 'html', 'latex', 'markdown', 'notebook', 'pdf',
    'python', 'rst', 'script', 'slides', 'webpdf'] or a dotted object name that
    represents the import path for an `Exporter` class
    Default: ''
    Equivalent to: [--NbConvertApp.export_format]
--template=<Unicode>
    Name of the template to use
    Default: ''
    Equivalent to: [--TemplateExporter.template_name]
--template-file=<Unicode>
    Name of the template file to use
    Default: None
    Equivalent to: [--TemplateExporter.template_file]
--writer=<DottedObjectName>
    Writer class used to write the  results of the conversion
    Default: 'FilesWriter'
    Equivalent to: [--NbConvertApp.writer_class]
--post=<DottedOrNone>
    PostProcessor class used to write the results of the conversion
    Default: ''
    Equivalent to: [--NbConvertApp.postprocessor_class]
--output=<Unicode>
    overwrite base name use for output files. can only be used when converting
    one notebook at a time.
    Default: ''
    Equivalent to: [--NbConvertApp.output_base]
--output-dir=<Unicode>
    Directory to write output(s) to. Defaults to output to the directory of each
    notebook. To recover previous default behaviour (outputting to the current
    working directory) use . as the flag value.
    Default: ''
    Equivalent to: [--FilesWriter.build_directory]
--reveal-prefix=<Unicode>
    The URL prefix for reveal.js (version 3.x). This defaults to the reveal CDN,
    but can be any url pointing to a copy  of reveal.js.
    For speaker notes to work, this must be a relative path to a local  copy of
    reveal.js: e.g., "reveal.js".
    If a relative path is given, it must be a subdirectory of the current
    directory (from which the server is run).
    See the usage documentation
    (https://nbconvert.readthedocs.io/en/latest/usage.html#reveal-js-html-
    slideshow) for more details.
    Default: ''
    Equivalent to: [--SlidesExporter.reveal_url_prefix]
--nbformat=<Enum>
    The nbformat version to write. Use this to downgrade notebooks.
    Choices: any of [1, 2, 3, 4]
    Default: 4
    Equivalent to: [--NotebookExporter.nbformat_version]

Examples
--------

    The simplest way to use nbconvert is
    
    > jupyter nbconvert mynotebook.ipynb --to html
    
    Options include ['asciidoc', 'custom', 'html', 'latex', 'markdown', 'notebook', 'pdf', 'python', 'rst', 'script', 'slides', 'webpdf'].
    
    > jupyter nbconvert --to latex mynotebook.ipynb
    
    Both HTML and LaTeX support multiple output templates. LaTeX includes
    'base', 'article' and 'report'.  HTML includes 'basic' and 'full'. You
    can specify the flavor of the format used.
    
    > jupyter nbconvert --to html --template lab mynotebook.ipynb
    
    You can also pipe the output to stdout, rather than a file
    
    > jupyter nbconvert mynotebook.ipynb --stdout
    
    PDF is generated via latex
    
    > jupyter nbconvert mynotebook.ipynb --to pdf
    
    You can get (and serve) a Reveal.js-powered slideshow
    
    > jupyter nbconvert myslides.ipynb --to slides --post serve
    
    Multiple notebooks can be given at the command line in a couple of 
    different ways:
    
    > jupyter nbconvert notebook*.ipynb
    > jupyter nbconvert notebook1.ipynb notebook2.ipynb
    
    or you can specify the notebooks list in a config file, containing::
    
        c.NbConvertApp.notebooks = ["my_notebook.ipynb"]
    
    > jupyter nbconvert --config mycfg.py

To see all available configurables, use `--help-all`.

[NbConvertApp] WARNING | pattern 'EL HALABI_KitDataScience_Projet Final.ipynb' matched no files